Legal Requirements:
For AMD software being released as open source, add copyright at the top of each new file.
Signed-off-by: Alysa Liu <Alysa.Liu@amd.com>
* Fix grid_group::group_dim to return grid_dim and not block_dim
* Add unit test for grid_group.group_dim()
* Fix unit test errors
* Skip group_dim() assertions for base_type test
* Added Fortran (amdflang) openmp tests using the openmp-vv project
---------
Signed-off-by: David Galiffi <David.Galiffi@amd.com>
Co-authored-by: David Galiffi <David.Galiffi@amd.com>
* add more derived metrics for navi4.
* addr comments
* addr comments, and add more derived counters.
* EOF.
* misc.
* remove duplicate counter.
* misc.
* Remove gfx12 architecture definition for ldslatency
* remove extra architectures for gfx12.
* use wgp for normalization
* move these changes to another PR.
---------
Co-authored-by: Venkateshwar Reddy Kandula <venkateshwar.kandula1306@gmail.com>
If a compressed changelog exists from a previous build, reconfiguring
the project fails with
```
[rocm-core configure] CMake Error at utils.cmake:213 (message):
[rocm-core configure] Failed to compress: gzip:
[rocm-core configure] /home/ben/src/TheRock/build/base/rocm-core/build/DEBIAN/changelog.Debian.gz
[rocm-core configure] already exists; not overwritten
```
Add `-f` to force overwriting.
* The QueuePair object was out of scope at the end of the for loop.
So the deconstructor was called.
* Although correct for C++ to do this, it ignores that we copied the QueuePair object into
device memory and have an instance there.
* Early deconstruction resulted in calling ibv_dereg_mr on the atomics memory region.
So when the GPU kernel tried to use the memory region it wasn't
registered which resulted in a protection domain error.
* The solution was to allocate our QueuePair obj with the new operator which leaves memory
management to us, then we can manually call the deconstructor.
[ROCm/rocshmem commit: e856fbb0eb]
* The QueuePair object was out of scope at the end of the for loop.
So the deconstructor was called.
* Although correct for C++ to do this, it ignores that we copied the QueuePair object into
device memory and have an instance there.
* Early deconstruction resulted in calling ibv_dereg_mr on the atomics memory region.
So when the GPU kernel tried to use the memory region it wasn't
registered which resulted in a protection domain error.
* The solution was to allocate our QueuePair obj with the new operator which leaves memory
management to us, then we can manually call the deconstructor.
* [SWDEV-531904] Unit and Integ Test Updates
Updated: unit_tests.py
- Removed redundant self.setUp() and self.tearDown() calls.
- Removed test_free_name_value_pairs() since is internal only.
Updated: integration_test.py
- Added logic to set AMDSMI_CLI_PATH from environment or default.
- Raise FileNotFoundError if path does not exist.
- Append CLI path to sys.path and handle ImportError with a clear message.
- Removed redundant @handle_exceptions function decorator.
- Removed redundant self.setUp() and self.tearDown() calls.
Updated: amdsmi_interface.py
- Removed POINTER conversion in amdsmi_get_gpu_pm_metrics_info() and amdsmi_get_gpu_reg_table_info()
All tests pass/skip
Signed-off-by: Juan Castillo <juan.castillo@amd.com>
* Update tests/python_unittest/integration_test.py
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Castillo, Juan <Juan.Castillo@amd.com>
* Review Update 1
Modified: integration_test.py
- Added logic to properly loop through firmware list and display each name and version
Signed-off-by: Juan Castillo <juan.castillo@amd.com>
* Skip xgmi_err tests + improve running output
Changes:
1. Now check for elevated permissions
2. Skip xgmi_error related SYSFS tests, refer to xgmi_read_write.cc
(both are skipped)
3. Added list of tests and provided a summary of additional output
provided
Change-Id: Iefc85c270faad89c625e2bd7af397d24faed2437
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
---------
Signed-off-by: Juan Castillo <juan.castillo@amd.com>
Signed-off-by: Castillo, Juan <Juan.Castillo@amd.com>
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Co-authored-by: Juan Castillo <juan.castillo@amd.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Charis Poag <Charis.Poag@amd.com>
* [SWDEV-531904] Unit and Integ Test Updates
Updated: unit_tests.py
- Removed redundant self.setUp() and self.tearDown() calls.
- Removed test_free_name_value_pairs() since is internal only.
Updated: integration_test.py
- Added logic to set AMDSMI_CLI_PATH from environment or default.
- Raise FileNotFoundError if path does not exist.
- Append CLI path to sys.path and handle ImportError with a clear message.
- Removed redundant @handle_exceptions function decorator.
- Removed redundant self.setUp() and self.tearDown() calls.
Updated: amdsmi_interface.py
- Removed POINTER conversion in amdsmi_get_gpu_pm_metrics_info() and amdsmi_get_gpu_reg_table_info()
All tests pass/skip
Signed-off-by: Juan Castillo <juan.castillo@amd.com>
* Update tests/python_unittest/integration_test.py
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Castillo, Juan <Juan.Castillo@amd.com>
* Review Update 1
Modified: integration_test.py
- Added logic to properly loop through firmware list and display each name and version
Signed-off-by: Juan Castillo <juan.castillo@amd.com>
* Skip xgmi_err tests + improve running output
Changes:
1. Now check for elevated permissions
2. Skip xgmi_error related SYSFS tests, refer to xgmi_read_write.cc
(both are skipped)
3. Added list of tests and provided a summary of additional output
provided
Change-Id: Iefc85c270faad89c625e2bd7af397d24faed2437
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
---------
Signed-off-by: Juan Castillo <juan.castillo@amd.com>
Signed-off-by: Castillo, Juan <Juan.Castillo@amd.com>
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
Co-authored-by: Juan Castillo <juan.castillo@amd.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Charis Poag <Charis.Poag@amd.com>
[ROCm/amdsmi commit: 67eb541c15]
abstract out the usage of direct verbs functionality to use tables with
the function pointers. This will allow in a second step the library to
be simultaniously be compiled for multiple NIC vendors/DV libraries and
interfaces. For now, the conversion has been done for IB MLX5 and BCOM
DV, the Pensando AINIC is to follow soon.
[ROCm/rocshmem commit: 99b753f103]
abstract out the usage of direct verbs functionality to use tables with
the function pointers. This will allow in a second step the library to
be simultaniously be compiled for multiple NIC vendors/DV libraries and
interfaces. For now, the conversion has been done for IB MLX5 and BCOM
DV, the Pensando AINIC is to follow soon.
DISABLE_GET removed from ALL, idea is that the CI scripts will invoke a
subset that is known to work rather than ALL
[ROCm/rocshmem commit: 5dc7d4539e]
Alysa Liu
Danylo Lytovchenko
vstojilj
harkgill-amd
systems-assistant[bot]
Saeed, Oosman
Dmitrii
xuchen-amd
Venkateshwar Reddy Kandula
ycui1984
Julian Jose
Aryan Salmanpour
ammallya
amd-hsivasun
marandje
Aurelien Bouteiller
Kian Cossettini
habajpai-amd
abchoudh-amd
Marius Brehler
Yiltan
3049ac537468bd90fe07f2cbb3d7a83e_amdeng
Edgar Gabriel
Arm Patinyasakdikul
Jatin Chaudhary
Pengda Xie
jamessiddeley-amd
mberenjk
Wenkai Du